Hello, I'm new to this list and I have a question that I hope is
appropriate to ask here. Here at SIL International we've developed a
rendering engine called Graphite that can do all kinds of rendering for
complex writing systems. The system is extensible in that the rendering
behavior is defined by a program written in a language we've invented
called GDL (Graphite Description Language). This is important for our
purposes because it provides rendering capability for the many little-known
languages around the world that are not supported by major software
development efforts.

We're considering the possibility of integrating Graphite rendering into
Mozilla, which would provide a means for communication of all kinds of data
from the lesser-known languages of the world via a web browser. Right now
we're trying to evaluate the feasibility of this idea. The most important
aspects to this would be finding places in the code to hook into in order
to do the following:

(1) recognize that we need to render with Graphite (probably by means of
the font name),

(2) call the Graphite engine to perform the rendering (write the glyphs
onto the screen), and

(3) define copy operations. This is complicated by the fact that there is
not a one-to-one correspondence between the rendered glyphs and the
characters to be copied, and the information about the correspondence is
hidden in the Graphite engine and related data structures. We could
probably simplify the behavior if needed. This item would not be absolutely
essential for the first phase of effort, but would be nice.

I have not looked at the Mozilla code at all, yet. But would anybody have
an off-the-top-of-your-head idea of what kind of effort this would involve,
say in terms of person-months? (Of course this is assuming the right kind
of interface(s) on the Graphite engine; we may also need to rework things
somewhat from our end.)

I see from the web site that it is also possible to grab the Gecko layout
code to incorporate into our own stuff. It occurred to me that a better way
to go might be to build a plugin that uses Gecko. Am I right in assuming
that this would give us the XML/HTML/CSS support that we would like to
have? Any sense of which would be the easier approach?

Sorry for the rather naive questions, I'm just trying to get an initial
feel to try to evaluate the feasibility of some different options. Thanks
for any help you can give.

Sharon Correll
SIL International


Reply via email to